Towards more ecologically realistic scenarios of plant uptake modelling for chemicals: PAHs in a small forest

Science of the Total Environment 505 (2015) 329–337

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Towards more ecologically realistic scenarios of plant uptake modelling for chemicals: PAHs in a small forest Elisa Terzaghi a,b, Gabriele Zacchello a, Marco Scacchi a, Giuseppe Raspa c, Kevin C. Jones d, Bruno Cerabolini b, Antonio Di Guardo a,⁎ a

Department of Science and High Technology, University of Insubria, Via Valleggio 11, 22100 Como, Italy Department of Theoretical and Applied Sciences, University of Insubria, Via J. H. Dunant 3, 21100 Varese, Italy Department of Chemical Engineering, Materials, and Environment, “La Sapienza” University, Via Eudossiana 18, 00184 Rome, Italy d Lancaster Environment Centre, Lancaster University, Lancaster LA1 4YQ, UK b c

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• PAH air concentrations are affected by meteorological and ecological parameters. • SLA is directly proportional to PAH uptake rate of vegetation. • LAI needs to be split to know the importance of each species in PAH accumulation. • Attention must be paid when collecting deciduous species leaves over time. • A dynamic fate model is required if the data are to be interpreted reliably.

a r t i c l e

i n f o

Article history: Received 19 August 2014 Received in revised form 24 September 2014 Accepted 30 September 2014 Available online xxxx Editor: D. Barcelo Keywords: Environmental risk assessment Multimedia fate models Dynamic Plants Leaf traits PAHs

a b s t r a c t The importance of plants in the accumulation of organic contaminants from air and soil was recognized to the point that even regulatory predictive approaches now include a vegetation compartment or sub-compartment. However, it has recently been shown that many of such approaches lack ecological realism to properly evaluate the dynamic of air/plant/soil exchange, especially when environmental conditions are subject to sudden variations of meteorological or ecological parameters. This paper focuses on the development of a fully dynamic scenario in which the variability of concentrations of selected chemicals in air and plant leaves was studied weekly and related to the corresponding meteorological and ecological parameters, to the evaluate their influence. To develop scenarios for modelling purposes, two different sampling campaigns were performed to measure temporal variability of: 1) polycyclic aromatic hydrocarbon (PAH) concentrations in air of a clearing and a forest site, as well as in leaves of two broadleaf species and 2) two important leaf and canopy traits, specific leaf area (SLA) and leaf area index (LAI). The aim was to evaluate in detail how the variability of meteorological and ecological parameters (SLA and LAI) can influence the uptake/release of organic contaminants by plants and therefore air concentrations. A principal component analysis demonstrated how both meteorological and ecological

⁎ Corresponding author. Tel.: +39 031 238 6480; fax: +39 031 2386449. E-mail addresses: [email protected] (E. Terzaghi), [email protected] (G. Raspa), [email protected] (K.C. Jones), [email protected] (B. Cerabolini), [email protected] (A. Di Guardo).

http://dx.doi.org/10.1016/j.scitotenv.2014.09.108 0048-9697/© 2014 Elsevier B.V. All rights reserved.

330

E. Terzaghi et al. / Science of the Total Environment 505 (2015) 329–337

parameters jointly influence PAH air concentrations. SLA, LAI, as well as leaf density were showed to change over time and among species and to be directly proportional to leaf/canopy uptake rate. While hazelnut had the higher leaf uptake rate, maple became the most important species when considering the canopy uptake rate due to its higher LAI. Other species specific traits, such as the seasonal variation in production of new leaves and the timing of bud burst, were also shown to influence the uptake rate of PAHs by vegetation. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Plants play an important role in influencing the environmental fate of organic compounds which are taken up from both air and soil (Collins et al., 2006). A number of models with different degrees of complexity have been developed to describe the uptake and accumulation of organic chemicals in vegetation. They range from simple models which consider only one species (Trapp et al., 1994), leaves as the unique compartment (Priemer and Diamond, 2002) and can be steady state models (Paterson et al., 1991), to more complex ones which consider mixed forests (Wania and McLachlan, 2001), different subcompartments (roots, stem, fruits and leaves) (Behrendt and Brüggemann, 1993) also in dynamic versions (Fantke et al., 2011). Plants are also included in the European Union System for the Evaluation of Substances (EUSES) (EC, 2004) which is a steady state model recommended in the European Union for risk assessment of organic chemicals. Recently it has been pointed out (Di Guardo and Hermens, 2013; EC, 2013) that the current procedure employed for environmental risk assessment lacks ecological realism and should be improved in the coming years. Concerning the exposure assessment, this can be achieved by development of dynamic multimedia fate models which are capable of predicting variable concentrations in time and space; however, in order to run such models a number of realistic scenarios should be developed to reflect ecosystem complexity, together with the temporal and spatial variability of environmental parameters. However, meteorological and compartment parameters currently used in multimedia fate models are generally based on average, annual or sometimes seasonal conditions, since realistic datasets (accounting for variation in time and space) are still scarce. Additionally, no models include a fully dynamic vegetation compartment, ignoring the variability of some leaf and canopy traits, such as specific leaf area (SLA) and leaf area index (LAI) as well as some species-specific behaviour such as the seasonal variation of production of new leaves or the timing of budburst, which all may have a role in influencing organic contaminant air concentrations. Additionally, employing this kind of dataset to predict the environmental fate of organic contaminants may lead to erroneous results or inconsistencies when comparing predictions to measured values. The aim of the present work was to evaluate in detail how the variability of a number of meteorological and ecological parameters (SLA and LAI) can influence the uptake/release of organic contaminants by plants and therefore air concentrations. SLA is a parameter that expresses the amount of leaf surface (one face) of a certain species per unit of dry weight (cm2/g or m2/kg) and it can be regarded as a measure of the surface available for organic pollutant exchange with air per units of leaf mass (Nizzetto et al., 2008). It is usually measured for fully expanded leaves representing SLA of the stable phase and only a few studies, including the present work, have focused on SLA determination over time (Bayrak Ozbucak et al., 2011; Luo et al., 2011; Nizzetto et al., 2007; Simioni et al., 2004). LAI is a measure of canopy foliage content defined as the amount of one side leaf area (m2) in a canopy per unit ground area (m2) (Asner et al., 2003) and it can be considered an index of the total foliar biomass accumulating organic pollutants from the air and a measure of the density of surface available for exchange with the atmosphere (Nizzetto et al., 2006). LAI and SLA allow calculation of a dynamic leaf biomass (B) since B = LAI/SLA. Since LAI refers to a whole canopy rather than to a single leaf or single species (such as SLA), in order to calculate the

role of the biomass change of the different species (e.g. to find out which ones are the most important) in driving the uptake, it is necessary to devise a way to split LAI into species contributions. The specific objective of this paper is the development of a fully dynamic scenario that considers the variability of exposure concentrations, meteorological and ecological compartment parameters. This scenario will be later adopted in a dynamic bioaccumulation model (SoilPlusVeg) that will be presented in a companion paper (Terzaghi et al., in preparation) to predict the temporal uptake and release (on an hourly basis) of some polycyclic aromatic hydrocarbons (PAHs) in a mixed broadleaf forest located in Northern Italy. In order to develop the scenario, two different sampling campaigns were conducted to measure variations over time of: 1) PAH concentrations in air of a forest and a clearing site, as well as in leaves of two different species and 2) two important ecological parameters (SLA and LAI). A number of meteorological parameters (wind speed, temperature, precipitations, planetary boundary layer, etc.) were also obtained. 2. Materials and methods 2.1. Sampling site Sampling was carried out in a small forest of about 7 ha located in the surroundings of the Department of Science and High Technology (Dept. SHT) of the University of Insubria in Como, a 60,000 inhabitant town located in Northern Italy. The sampled area could be classified as an urban background site according to Larssen et al. (1999) due to its distance from the nearest street (about 150 m). This site was selected because it allowed the collection of samples in a small deciduous forest and in an adjacent clearing area (100 m away), which consisted of a meadow with sporadic trees. 2.2. Sampling and analysis Two different sampling campaigns were performed: 1. From March to July 2007 leaves of two tree species (cornel, Cornus mas and maple, Acer pseudoplatanus) and air samples were collected for PAH measurements; 2. From March to December 2012 leaves of the previous species plus an additional one, hazelnut (Corylus avellana), were sampled for SLA and LAI determinations. At the same time leaf development of the three species was followed employing a non-destructive method to understand the leaf area variability range. In August the forest was surveyed (August 1st, 2012) to measure canopy composition. PAH analysis was performed for: acenaphthylene (acy), acenaphthene (ace), fluorene (fluo), phenanthrene (phe), anthracene (anth), fluoranthene (flout), pyrene (pyr), banzo[a]anthracene (b[a]anth) chrysene (chr), benzo[b]fluoranthene (b[b]flout), benzo[a]pyrene (b[a]pyr), perylene (per), benzo[ghi]perylene (b[ghi]per), dibenzo[ah]anthracene (di[ah]anth), indeno[cd]pyrene (i[cd]pyr) and coronene (cor) following the analytical procedure reported in Terzaghi et al. (2013). Supporting information (SI) (Text SI-A1-A7) report details about air (Text SI-A1) and leaf sampling (Text SI-A2) for PAH determination, PAH analysis and Quality Assurance/Quality Control (QA/QC) (Text SI-A3), SLA (Text SI-A4) and LAI measurement (Text SI-A5), the nondestructive method employed to follow leaf development (Text SI-A6) and forest composition determination at full growth (Text SI-A7). Briefly, air sampling was performed with Hi-vol sampler separately analysing gas and particle phase. Leaves (about 100 g) were sampled

E. Terzaghi et al. / Science of the Total Environment 505 (2015) 329–337

weekly during the sampling campaign. PAHs were extracted from PUFs (gas phase), GFF (particle phase) and leaves using dichloromethane and purified following a gel permeation (only leaves) and silica–alumina cleanup procedure. Extracts were analysed in GC-MS in single ion mode using deuterated standards as internal standards. SLA was measured according to Cornelissen et al. (2003) taking 10 leaves per sampling day for each species. LAI was measured using a LI-COR LAI-2000 Plant Canopy Analyzer. More details are reported in Text SI-A1 to 7. 2.3. Meteorological parameters Measurements of air temperature were carried out with a Testo temperature logger (Testo AG, Lenzkirch, Germany, mod. 174) in the clearing and the forest site. Loggers were sheltered to protect them from direct sunlight and precipitations. Temperature was measured at 1 hour intervals for the entire sampling period. Other meteorological parameters such as precipitation and wind speed were obtained from the Regional Environmental Protection Agency (REPA, 2014) for Como city. Hourly lower air heights (height of planetary boundary layers, PBL) were obtained as described in Morselli et al. (2011). 2.4. Statistical analysis Principal component analysis (PCA) was used to investigate the joint influence of meteorological and ecological parameters on air and leaf PAH concentrations. Only cornel concentrations (phe, fluoth, pyr) were used in PCA, since the dataset was complete (all weeks of sampling) and comparable to the other variables. PCA was performed with a programme developed by one of the authors according to Saporta (1990). 3. Results and discussion 3.1. Overview of the results obtained

331

and SLAh (SLA of hazelnut). Rainfall and wind speed are representative of the second component (PC2) and in an opposite way of the third principal component (PC3). The correlation matrix among the variables can be seen in Table SI-A14. The association of temperature and PBL height to the ecological parameters representative of the whole forest, such as leaf biomass and LAI can be explained considering that those variables are related to the conditions responsible for the development of the forest (Larcher, 2003). Concerning the other ecological parameters which depend on plant species, such as SLA, a different behaviour can be noted: while maple SLA is positively associated to temperature, PBL height, LAI and leaf biomass, SLA of both cornel and hazelnut are correlated in an opposite way to all these parameters. This could be due to a different type of leaf development for the three plant species, in response to meteorological parameters. Cornel and hazelnut leaf biomass appear about three weeks before maple while temperature starts to grow. Maple foliar development starts at a later time and being much faster, rapidly dominates total surface area (see 3.3). It is well known that SLA is influenced by the availability of light during leaf growth (Poorter and Evans, 1998): cornel and hazelnut, which are under storey species, have to rapidly expand their leaf area to capture the light for photosynthesis before that maple, which is an upper canopy species, covers them. Rain and wind are little correlated with each other and also to the other meteorological and ecological variables. They are in agreement with PC2 while they are contrastingly but highly correlated on PC3. This suggests that PC2 and PC3 could represent two different meteorological situations present in some of the weeks. The above mentioned 9 variables were then plotted over the planes PC1 to PC2 and PC1 to PC3 and the supplementary variables (PAH air and leaf concentrations) were over imposed (Fig. 3.1 and Figure SI-A1, with loadings in Table SI-A15). From Fig. 3.1 (PC1 vs. PC2) it appears that PAH concentrations in air are generally opposed (with a varying degree), in respect to the PC1 axis, to temperature, PBL height, LAI and leaf biomass, indicating that all these parameters (including reduced emissions in summer) contribute to PAH air concentration reduction (Nizzetto et al., 2006; 2008; Morselli et al., 2011). A few groups appear:

Complete tables reporting the measured ecological parameters (SLA and LAI), PAH concentrations measured in air and leaves and meteorological parameters are in the SI (Table SI-A2–A12). These data were used for a preliminary and exploratory analysis using PCA, while later the influence of the temporal variations of ecological and meteorological parameters was investigated to explore their influence on PAH concentrations measured in air (clearing and forest) and leaves. Some chemicals (b[b]fluot, b[a]pyr, per, i[cd]pyr, db[ah]anth, b[ghi]per and cor) were never or seldom found in air samples, while only phe, fluot, pyr, and chr were generally detected in leaves. For maple leaves, concentrations were not available for three sampling weeks: the first two because at that time maple leaves were not present and the ninth because the sample was lost. The air samples of the period May, 24– May, 31 were not considered due to a problem with the Hi-vol sampler of the clearing site. 3.2. Principal component analysis The PCA was employed to investigate the joint influence of 9 variables: 4 meteorological (temperature, mixing layer height, rainfall, wind speed) and 5 ecological parameters (leaf biomass, LAI, and SLA of the three species) on gas and particle phase PAH concentrations, present as supplementary variables, both in clearing and forest air. In the present analysis the first three principal components (PC1, PC2, and PC3) accounted for 65%, 20% and 8% respectively, which cooperatively explained 93% of the total variance. The correlations of the 9 variables with the first three principal components are reported in Table SI-A13. The first component (PC1) is highly correlated with all meteorological and ecological parameters, with the exception of rainfall and wind speed, and in particular shows the antagonism of temperature, PBL, SLAm (SLA of maple), LAI and biomass to SLAc (SLA of cornel)

Fig. 3.1. Loading plot of PC1 and PC2. The group association is explained in the main text. (GC: gas phase clearing (pink); GF: gas phase forest (purple); PC: particle phase clearing (orange); PF: particle phase forest (brown); T: temperature, Rain: rainfall, Wind: wind speed, PBL: planet boundary layer height, SLAc: specific leaf area of cornel, SLAh: specific leaf area of hazelnut, SLAm: specific leaf area of maple, LAI: leaf area index, BIO: biomass; L: Leaf; PM10: particulate matter). For chemical abbreviations see the main text.

332

E. Terzaghi et al. / Science of the Total Environment 505 (2015) 329–337

“Group 1” includes compounds measured in the particle phase in forest air, with the exception of per while chemicals adsorbed to particles in the clearing air with the exception of per and fluot belong to “Group 2”. These two groups are separated by PC2 (mostly wind speed, associated to rain). They are basically the PAHs in air particles in the forest (negatively associated to wind speed, in other words to less windy conditions) and particle associated chemicals in the clearing (positively associated to wind speed). By looking at these differences, the role of the forest in capturing particle associated PAHs is evident (Terzaghi et al., 2013) and this is probably due to the reduction of wind speed within the canopy, as well as the role of leaf “sticky” surfaces in trapping/ retaining compounds. “Group 3” is spread in the PC1 to PC2 plane and comprises most of the gas phase chemicals in clearing and forest air, on which the influence of wind and rainfall is variable. The decrease from positive to negative PC2 seems to be generally related to increasing Log KOW of the chemicals. PAH leaf concentrations are positively correlated with temperature, PBL height, leaf biomass and LAI (phe and pyr show higher correlation than flout) and therefore negatively correlated with PAH air concentrations, again showing the importance of vegetation and PBL in reducing air concentration. However, since T, LAI, PBL height, and leaf biomass are strongly positively correlated (Table A14) they jointly influence the air concentration reduction.

3.3. Temporal variation of ecological parameters Ecological parameters were measured in 2012, five years later than the PAH determination in air and leaves; therefore the beginning of the 2012 growing season was shifted (see Text SI-A8) to match the leaf development start of 2007.

3.3.1. Specific leaf area Fig. 3.2(a) shows the temporal trend of SLA for cornel, maple and hazelnut measured during the present study. Table SI-A8–A10 reports leaf parameters used for the SLA calculation. SLA showed a similar trend for the three species. SLA presented an initial phase characterized by low values, but with a tendency to increase in the next one or two weeks. Young leaves are not fully hydrated and may have not fully turgid vacuoles. This affects the expansion of leaves and therefore influences the ratio of leaf area on the leaf dry weight resulting in a low SLA. In a few weeks, the hydration of vacuoles caused the full expansion of the young leaf represented by maximum SLA values. Finally, SLA decreased and became stable due to the building up of internal leaf structure and the development of the cuticular wax protective layer which caused leaf weight accumulation in time versus surface area expansion. A similar behaviour was found by Jurik (1986) In general, SLA values of the early stage of leaf development (end of March and beginning of April) are significantly different with respect to those of the stable phase (after the end of April). In terms of leaf uptake of chemicals, SLA can be regarded as a measure of the surface available for organic pollutant exchange with air per units of leaf mass (Nizzetto et al., 2008). As shown by McCrady (1994) and Steyaert et al. (2009), the uptake rate of organic pollutants to vegetation k1 (h−1) is directly proportional to the exposed surface area of the foliage according to the equation:

k1 ¼ kU

  A V

ð1Þ

where k1 is the uptake rate of organic pollutants to vegetation (h−1) on a volume/volume basis, kU is the mass transfer coefficient between the vegetation surface and the atmosphere (m/h), and A/V is the leaf area to volume ratio (m−1).

If the leaf area to volume ratio (A/V) is substituted with the leaf area to mass ratio (SLA) and leaf density, the following is obtained: k1 ¼ 2kU SLA ρL

ð2Þ

where k1 is the uptake rate of organic pollutants to leaves (h−1) on a volume/volume basis, kU is the mass transfer coefficient between the leaf and the atmosphere (m/h), SLA is the specific leaf area of the selected species (m2/kg) and ρL is the dry leaf density (kg/m3). The term “2” indicates that both sides of the leaf are considered. This means that SLA is directly proportional to the uptake rate k 1 (h− 1 ) of organic pollutants to vegetation; therefore lower SLA values will result in a slower increase in the leaf concentrations with time (Nizzetto et al., 2007; Nizzetto et al., 2008). As appears from Eq. (2), k1 depends not only on SLA but also from another leaf trait: ρL. Leaf density was shown to be positively correlated with leaf mass per area (LMA) and therefore negatively correlated with SLA (that is the inverse of LMA) (Castro Diez et al., 2000). According to Vile et al. (2005) the leaf dry matter content (LDMC) can be used as a surrogate of dry leaf density considering that leaf fresh mass is a good estimate of leaf volume. Therefore: LDMC ¼

Leaf dry mass ¼ ρL Leaf fresh mass

ð3Þ

where LDMC is the leaf dry matter content (kg/kg), leaf dry mass is expressed in kg, while leaf fresh mass is converted to leaf volume assuming a leaf fresh density ≈ 1000 kg/m3. Using these approximations, dry leaf density trends for the three species together with the uptake rate, calculated employing Eq. (2), can be obtained (Fig. 3.2 (b, c). A constant mass transfer coefficient (kU) of 9 m/h was assumed to standardize the comparison, although kU can also change with wind speed (Barber et al., 2004). Leaf density increases with time, while the uptake rate increases at the beginning (due to the increase of both SLA and ρL) and then remains quite stable (since ρL increases while SLA decreases of a similar amount). It is evident that hazelnut has the higher uptake rate, followed by cornel and maple. SLA values presented here were calculated for leaves characterized by an area always larger than those collected at the previous sampling date (Text SI-A4). The area of leaves collected for SLA determination seems to represent maximum leaf area available at the time of sampling. However, results of the non-destructive sampling show a high variability in leaf area (and generally lower values), probably representing the fact that different populations of leaf age (and area) were sampled. This small forest can be described as a variable combination of leaves which change their characteristics (leaf area, SLA, leaf density) with time but not in a synchronous way (see also Figure SI-A4). Such an effect should be taken into account when sampling leaves for SLA or chemical contamination reasons. 3.3.2. Leaf area index As mentioned in the introduction, LAI can be regarded as an index of the total foliar biomass accumulating organic pollutants from the air and a measure of the density of surface available for exchange with the atmosphere (Nizzetto et al., 2006). Table SI-A11 reports LAI values for the mixed wood of the present study. Around the beginning of July LAI reached a maximum of about 4.5 and then dropped due to litterfall, reaching the initial value of about 0.70. Unlike SLA, LAI is an ecological parameter which refers to a whole canopy rather than to a single leaf or single species. As the knowledge of the contribution of each species to the total LAI could allow calculating a species specific leaf biomass, in the present study an attempt to “distribute” LAI was made (Text SI-A9 and Fig. 3.3). Such specific LAI would allow, for example, to simulate the effect of single species on chemical uptake, or to assemble a modelling scenario of different forests. As shown in Fig. 3.3 the most important contribution to total LAI was given by maple,

E. Terzaghi et al. / Science of the Total Environment 505 (2015) 329–337

333

Bud burst 50 End of cornel leaf sampling 2 SLA (m /kg)

45 40 35 30 25

a

Maple

Cornel

06-Oct

22-Sep

08-Sep

25-Aug

11-Aug

28-Jul

14-Jul

30-Jun

16-Jun

02-Jun

19-May

05-May

21-Apr

07-Apr

24-Mar

10-Mar

25-Feb

20

Hazelnut

Bud burst

3

Leaf density (g/cm )

0.45 0.40 0.35 0.30 0.25 0.20 0.15

End of cornel leaf sampling

b

Maple

Cornel

06-Oct

22-Sep

08-Sep

25-Aug

11-Aug

28-Jul

14-Jul

30-Jun

16-Jun

02-Jun

19-May

05-May

21-Apr

07-Apr

24-Mar

10-Mar

25-Feb

0.10

Hazelnut

Bud burst 2.5E+05

Uptake rate (1/h)

2.2E+05 1.9E+05 End of cornel leaf sampling

1.6E+05 1.3E+05 1.0E+05

c

Maple

Cornel

06-Oct

22-Sep

08-Sep

25-Aug

11-Aug

28-Jul

14-Jul

30-Jun

16-Jun

02-Jun

19-May

05-May

21-Apr

07-Apr

24-Mar

10-Mar

25-Feb

7.0E+04

Hazelnut

Fig. 3.2. Specific leaf area (a), leaf density (b) and uptake rate (c) of cornel, hazelnut and maple (average ± standard error). Dates correspond to the sampling day of the leaf used for SLA determination, synchronized to match the leaf development start of 2007 (see Text A8 for details). Dotted lines represent bud-burst days of the three species.

followed by hazelnut and cornel, with the exception of the first two weeks when maple leaves were not present (maple bud burst occurred on April, 7 while that of hazelnut and cornel about three weeks before, on March, 15). This LAI distribution is useful to calculate the leaf biomass of the different species (Figure SI-A6), using SLA of each species. According to Barber et al. (2004), the mass transfer coefficient of a canopy of leaves kUcanopy can be calculated using kU of all the individual leaves in the canopy and relating to the total number of leaves (related to LAI). Therefore, Eq. (2) can be modified in: k1canopy ¼ 2ðkU LAIÞSLA ρL

ð4Þ

where k1canopy is the uptake rate of organic pollutants to the canopy (h−1) on a volume/volume basis, kU is the mass transfer coefficient between the canopy and the atmosphere (m/h), SLA is the specific leaf area of the selected species (m2/kg), ρL is the dry leaf density (kg/m3), and LAI is the leaf area index. The term “2” indicates that both sides of the leaf are considered. This equation combines two leaf traits (SLA and ρL) with a canopy trait (LAI) and it is useful to estimate the uptake rate of organic contaminants by the whole canopy evaluating the contribution of each species in removing chemicals from air. As shown in Fig. 3.4, when considering LAI, maple becomes the most important species (followed by hazelnut and cornel) in the uptake of contaminants from the air, driving the

334

E. Terzaghi et al. / Science of the Total Environment 505 (2015) 329–337

5

Cornel

Hazelnut

Maple

m2/m2

4

3

1.8

1.9

2.1

2.1

2.2

2.3

1.2 0.6

2

1

0

1.1

1.1

1.1

1.3

1.3

1.3

1.3

1.4

1.2

0.7

0.6

0.6

0.6

0.5

0.5

0.5

0.5

0.5

0.6

0.4 0.3

0.5

31

17

10

03

26

19

12

05

29

22

15

ay

ay

M

M

ay

ay

M

M

-7 ay

M

ay

M

ay

M

ay

M

pr

pr

A

A

n Ju

4

7

0

-1

-1

03

26

-2

pr

pr

A

A

19

pr

pr

A

ar

M

ar

M

A

9

5

12

-0

-2

2

-2

pr

ar

ar

pr

A

A

M

M

ar

M

Fig. 3.3. Specific leaf area index of maple, hazelnut and cornel (average values considering the air sampling weeks).

uptake rate since bud burst. This trend could not be appreciated from Fig. 3.2 where the uptake rate of single leaves of each species appeared and maple leaves seemed to be the least efficient in the uptake.

depending on the different thermal requirement (Wesolowski and Rowinski, 2006). Also within the same species, individual trees do not start leaf development synchronously. This behaviour was shown for maple and cornel and it is illustrated in Figure SI-A7 in which leaf area percentile variation with time is shown (see also 3.5). This aggregate behaviour shows that forests (especially Populustypes) should probably be described as a mixture of older and fresh leaves and the amount of fresh leaves appearing during canopy development should be estimated to accurately describe the uptake behaviour. This behaviour could be of course more or less important depending also on the length of the growing period, varying altitudinally and latitudinally.

3.3.3. Leaf production timing In evaluating the capability of a tree or a forest to uptake organic contaminants from air, another important aspect to be considered is the seasonal variation in production of new leaves of deciduous trees. The trees of the so-called Quercus-type (e.g. Quercus robur, Fagus sylvatica) produce new leaves mainly in spring, while new foliage in the trees of Populus-type (e.g. Populus spp., Betula pendula) are continuously produced up to the late growing season (Larcher, 2003). The small forest investigated during the present study is composed of two species characterized by “indeterminate leaf production” (or Populus-type) (C. mas and C. avellana) and another (A. pseudoplatanus) that although it is generally classified as a “single leaf flush” species (or Quercustype) (Nasahara et al., 2008) it could show an intermediate behaviour. In a forest mainly composed of Populus-type species, new leaf biomass could appear during the whole growing season: these new leaves could continuously deplete air thanks to their fresh organic carbon, high SLA and k1 (see 3.3.1). Therefore, different forests would differ for the amounts of fresh leaves appearing at each time periods (e.g. 15%), capable of capturing chemicals at the maximum air/plant gradient. Moreover, plant species can exhibit different timing of bud burst

3.4. Temporal variation of gaseous and particulate PAHs in clearing and forest air During each sampling period PAH levels in the particulate phase were lower than those in the gaseous phase by an average factor of about 10 and 8 for the clearing and the forest site respectively (Table SI-A2–A5). This observation was in agreement with previous studies (Choi et al., 2008; Dachs et al., 2002; Kishida et al., 2011). The most abundant chemical found in the gaseous phase of clearing and forest air was phe representing respectively 36–72% and 58–76% of total gas phase PAHs, while per dominated in the particulate phase of both

Bud burst

Canopy uptake rate (1/h)

4.2E+05 3.7E+05 3.2E+05 2.7E+05 2.2E+05 1.7E+05

End of cornel leaf sampling

1.2E+05 7.0E+04

Maple

Cornel

06-Oct

22-Sep

08-Sep

25-Aug

11-Aug

28-Jul

14-Jul

30-Jun

16-Jun

02-Jun

19-May

05-May

21-Apr

07-Apr

24-Mar

10-Mar

25-Feb

2.0E+04

Hazelnut

Fig. 3.4. Canopy uptake rate (1/h). Dates correspond to the sampling day of the leaf used for SLA determination, synchronized to match the leaf development start of 2007 (see Text A8 for). Dotted lines represent bud-burst days of the three species.

E. Terzaghi et al. / Science of the Total Environment 505 (2015) 329–337

sites, constituting 5–36% (clearing) and 6–32% (forest) of total particulate phase PAHs. Temporal trends show a fast decrease in total bulk and gas phase PAH levels for both the clearing and the forest sites (Fig. 3.5a), between March, 22 and April, 5. After this rapid decrease concentrations experience a less pronounced increase until April, 12 and remained quite stable until the end of the sampling period in the clearing site air, while they slowly but constantly were reduced in the forest air. Many factors may have affected PAH air concentrations during the whole sampling campaign. Fig. 3.5a shows PAH air concentrations in clearing and forest sites, together with some meteorological parameters trend (PBL height and temperature). As shown by the PCA (see 3.1) PAHs are negatively correlated with temperature and PBL height. The other two parameters (precipitation and wind speed) showed poor correlation with PAH concentrations. In order to understand the impact of the individual effects of different meteorological parameters on PAH air concentrations, an evaluation of the temporal contribution of the factors involved should be performed. Hornbuckle and Eisenreich (1996) reported that temperature controls volatilization and deposition of SOCs from/to surfaces (soil and vegetation) and therefore it may be responsible of the increase of SOC air concentrations during the day (or during period of high temperatures) and the decrease during the night (or during period of low temperatures). However, there are other processes, such as photo-degradation and mixing due to atmospheric turbulence, that could induce the opposite trend in SOC air concentration. Both photo-degradation and mixing due to atmospheric turbulence are positively correlated with daytime temperature (causing a reduction in PAH air concentrations). Moreover, for PAHs there are important

335

ongoing primary atmospheric sources that could be responsible of the temporal variability of their concentrations (Lee et al., 1998). If this is true the negative correlation between temperature and PAH air concentrations could be due to the competing influence of the other meteorological parameters (such as PBL height), as well as by recent local/ regional sources. PBL could have important diel variability, as shown in Morselli et al., 2011, and such influence could change air concentrations of a factor of 20 to 30. In the current study this can be observed after April, 26 when the increase of PBL height, together with OH radical presence and other solar induced reactions, might have caused a dilution and/or degradation of PAHs in air, thus masking the effect of temperature in enhancing the volatilization of contaminants from surfaces (soil and vegetation) and the consequent increase in PAH air concentrations. A similar behaviour was reported by Gu et al. (2010) who found a moderate negative linear relationship between temperature and PAH air concentrations. Meteorological parameters are not the sole controlling factors of the PAH air concentration variations. Other factors such as PAH emissions and the presence of vegetation can also influence PAH air concentrations. It is well known that the atmospheric origin of PAHs is very different: by-products from incomplete combustion of fossil fuels and wood, residential heating, coke production and vehicular traffic (Morville et al., 2011). Some authors (Cabrerizo et al., 2011; Wilcke, 2007) have also shown the biogenic sources of some PAHs from degradation of organic matter. Residential heating and vehicular traffic can be considered as a major source of PAHs in Como air (REPA, 2007). Unfortunately, data on the emission relevance for the different sources of PAHs in Como is not available; therefore the contribution of this factor in influencing

CA, clearing

PBL height (m)

2500

30 Temperature

CA, forest

2000 PBL height

25 20

1500 15 1000

10

500

5

0

0 ay M

ay M

ay M

ay M

-7

ay M

ay M

ay M

ay M

pr

A

pr

A

Cornel

n Ju

4 -2

7 -1

6

03

-2

9

-1 pr

pr

A

ar M

ar M

A

0 -1

pr A

pr

A

pr

A

05

29

22

2 -1

ar

pr

A

M

ar M

ar M

55

31

17

10

03

26

19

12

05

29

22

15

a

T (°C) - Air. conc. (ng/m3)

35

3000

Maple

50

Leaf Conc. (ng/g fw)

45 40 35 30 25 20 15 10 5 0

b

29-Mar

05-Apr

12-Apr

19-Apr

26-Apr

03-May

10-May

17-May

24-May

31-May

07-Jun

Fig. 3.5. a) Trend of temperature (°C) (red), PBL height (m) (blue), PAH bulk concentration in clearing (light blue) and forest (green) (ng/m3). Box plot refers to PBL height (90th, 75th, median, 25th, 10th). For temperature average values are plotted. b) PAH concentration (sum of phe, fluot, pyr, chr) in cornel and maple leaves. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

336

E. Terzaghi et al. / Science of the Total Environment 505 (2015) 329–337

air concentration cannot be evaluated. As mentioned above, the decrease in PAH air concentrations (March, 22 to April, 5) can be probably attributed to the appearance of “fresh uncontaminated vegetation”, acting as a filter for this type of compounds (McLachlan and Horstmann, 1998; Gouin et al., 2002; Nizzetto et al., 2006). As shown in Figure SIA6, leaf biomass showed an increase of more than 50% in 21 days, starting March, 15. In the two sites PAH air levels in the forest always resulted in lower concentrations than clearing concentrations, with the exception of the 2 weeks (April 5 to 19), when maple leaves are still at the beginning of their development. Before Mar 29, some vegetation was already present but probably leaf biomass was too low (LAI ~ 1 m2/m2) to influence air concentrations. Initial foliar development started on March, 15 (with the appearance of cornel and hazelnut leaves), while a second bloom appeared around April, 5–12 when maple leaves appeared very quickly, probably due to the rain events of the previous weeks followed by an increase of the average temperatures. However, during the following days of the growing season the increase of leaf biomass is responsible of the reduction and the stability of low PAH concentrations also in period characterized by low median mixing layer height (e.g. April, 12–April, 19 and April, 19–April, 26). Total particulate PAH trends presented some differences compared to the gas phase. Higher concentrations were found in the forest than in clearing site between March, 22 and April, 26. From March, 22 to April, 26 the sum of total particulate PAHs was about 8 and 12 ng/m3 for the clearing and the forest sites respectively. Higher concentrations in the forest were probably due 1) to bud burst and so to the loss of bud protective structure which passed the previous winter period and so were probably enriched by contaminants and/or 2) to biogenic particles release or anthropogenic particle erosion from leaves due to precipitation or wind (Terzaghi et al., 2013). Both bud protective structure and biogenic/anthropogenic particles may have acted as a sort of “scavenging” material, enhancing air particulate phase PAHs. 3.5. Temporal variation of PAHs in cornel and maple leaves The most abundant congener found in leaves was phe representing between 21–65% (cornel) and 19–47% (maple) of total PAHs (Table SI-A6–A7). Only were three congeners detected almost each sampling day: fluot, pyr and chr. They represented between 60–100% (cornel) and 70–100% (maple) of total PAHs, similar to atmospheric gaseous PAHs composition (48–92% in clearing and 67–88% in forest). Ace, acy, fluo, anth, b[a]anth, b[b]flout, b[a]pyr, per, ic[d]pyr, db[ah]anth, b[ghi] per, cor were never or just sometimes detected in leaves since they have log KOA lower than 7 or higher than 11 which are not in the expected range for a pronounced filter effect (McLachlan and Horstmann, 1998). Cornel and maple show different accumulation trend. PAH concentration in cornel (Fig. 3.5b) was high at the beginning of the season, due to the high air concentration, then decreased and remain quite stable (around an average value of 15 ng/g fw), excluding the last sampling day, when they reached a value of 27 ng/g fw (half of the initial concentration). Although the reduction in PAH air concentration was attributed to the forest filter effect, leaves do not seem to increase their concentration during six weeks of exposure. The quite constant concentrations found in both cornel and maple leaves during the period April, 12 to May, 24 can be attributed to different factors. First of all, leaves of quite different sizes for both maple and cornel were collected during the same sampling day (Figure SI-A7). Therefore each sample was a mixture of old and new leaves characterized by different length of exposure period as well as by different PAH concentrations. The analysis of this type of mixed sample resulted in a concentrations that substantially did not change with time. The important consequences of this fact are that when sampling leaves of broadleaf species, attention must be paid in selecting leaf samples, in order to collect leaves which started the development at the same time. Other factors which may have played a role in keeping constant the concentration in leaves are volatilization and

photo degradation on leaf surface enhanced by temperature increase. An additional factor which should be taken into account in the evaluation is the fact that PAH leaf concentrations are obtained at a single time of sampling, while air concentrations reflect the average value of a whole week. Sudden variations of air concentration during high and low PBL times (e.g. day and night) can be probably responsible for reverting the air/leaf gradient and therefore favouring the release or the uptake during those periods. If we compare the ratio of concentrations in leaves to concentrations in air (Cl/Ca) to Koa (Figure SI-A8) of the different (prevalently) vapour-phase chemicals (phe, fluot and pyr), we can observe that the ratio oscillates around the predicted leaf/air partition coefficient. From this appears that equilibrium is not reached at the end of the uptake period for these chemicals, showing the quick response of leaves to the change of air concentrations. This phenomenon cannot be completely accounted for in the present work because its evaluation requires the use of an integrated air/vegetation model and will be the subject of a companion paper (Terzaghi et al., 2013). 4. Conclusions The present study investigated how the variability of meteorological and ecological parameters can influence the uptake and release of PAHs by leaves and air concentrations. The uptake rate of PAHs by vegetation seemed to be influenced by both leaf (SLA, density) and canopy (LAI) traits as well as by species specific behaviour (seasonal variation in production of new leaves, timing of bud burst). To improve the ecological realism in exposure assessment, all these leaf and canopy traits and species behaviour should be considered when developing new dynamic multimedia fate models that include a vegetation compartment. Presently, when SLA is used in models, it is considered constant over time (Rein et al., 2011); leaf density is set to a fixed value as well as the timing of bud burst. The possibility for one species to produce new leaves also late in the season is not taken into account. Also LAI is a parameter that, when considered, is assumed constant over time, representing the value of the maturity phase (Bathia et al., 2008; Komprda et al., 2009; Priemer and Diamond, 2002; Undemann et al., 2009; Wegmann et al., 2004) of the canopy as a whole, while a specific LAI, associated with SLA to calculate leaf biomass, has never been employed. Leaves did not appear to accumulate PAH with time, probably due to volatilization or photo-degradation from/on leaf surfaces and the collection of a mixture of leaves characterized by different exposure time. This highlighted the need for a protocol for the sampling of deciduous leaf when the seasonal (or shorter time) variability of leaf contaminant concentrations has to be investigated. Furthermore current available datasets do not allow comparison of air and leaf fluxes at short time intervals (e.g. days or hours), therefore new measurements are needed. Other factors should also be taken into account: for example the stratified nature of leaves in woods which could be responsible for different photodegradation sensitivity of chemical in the different layers; or the role of falling leaves (e.g. litter) reaching and interacting with the soil ecosystem in releasing and distributing chemicals between air and soil compartments. In order to accurately predict air and plant concentrations in time, a dynamic multimedia fate model implementing the effect of the temporal variability of both ecological and meteorological parameter is therefore envisaged. Acknowledgement We thank Luca Nizzetto for the help in air sampling planning, Elisa Giovanniello for SLA and LAI measurement and Vito Falanga for the woods composition survey. Alberto Vianelli is also acknowledged for the advice regarding leaf density determination. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2014.09.108.

E. Terzaghi et al. / Science of the Total Environment 505 (2015) 329–337

References Asner GP, Scurlock JMO, Hicke JA. Global synthesis of leaf area index observations implications for ecological and remote sensing studies. Glob Ecol Biogeogr 2003;12: 191–205. Barber JL, Thomas GO, Kerstiens G, Jones KC. Current issues and uncertainties in the measurement and modelling of air-vegetation exchange and within-plant processing of POPs. Environ Pollut 2004;128:99–138. Bathia MA, Kadhum AAB, Mohamad AB, Takriff MS, Fisal Z, Daud WRD, et al. MAM—an aquivalence-based dynamic mass balance model for the fate of non-volatile organic chemicals in the agricultural environment. Am J Appl Sci 2008;1:252–9. Bayrak Ozbucak T, Kutbay HG, Yalcin S, Kilic DD. Foliar nitrogen (N), phosphorus (P) dynamics, and foliar resorption on Corylus avellana var. avellana. Ekoloji 2011; 20:1–7. Behrendt H, Brüggemann R. Modelling the fate of organic chemicals in the soil plant environment: model study of root uptake of pesticides. Chemosphere 1993;27(12): 2325–32. Cabrerizo A, Dachs J, Moeckel C, Ojeda M-J, Caballero G, Barceló D, et al. Ubiquitous net volatilization of polycyclic aromatic hydrocarbons from soils and parameters influencing their soil–air partitioning. Environ Sci Technol 2011;45:4740–7. Castro-Diez P, Puyravaud JP, Cornelissen JHC. Leaf structure and anatomy as related to leaf mass per area variation in seedlings of a wide range of woody plant species and type. Oecologia 2000;124:476–86. Choi SD, Staebler RM, Li H, Su Y, Gevao B, Harner T, et al. Depletion of gaseous polycyclic aromatic hydrocarbons by a forest canopy. Atmos Chem Phys 2008;8:4105–13. Collins C, Fryer M, Grosso A. Plant uptake of non-ionic organic chemicals. Environ Sci Technol 2006;40:45–52. Cornellissen JHC, Lavorel S, Garnier E, Diaz S, Buchmann N, Gurvich DE, et al. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust J Bot 2003;51:335–80. Dachs J, Glenn IV TR, Gigliotti CL, Brunciak P, Totten LA, Nelson ED, et al. Processes driving the short-term variability of polycyclic aromatic hydrocarbons in the Baltimore and northern Chesapeake Bay atmosphere, USA. Atmos Environ 2002;36:2281–95. Di Guardo A, Hermens JLM. Challenges for exposure prediction in ecological risk assessment. Integr Environ Assess Manag 2013;9:4–14. EC. European Union System for the Evaluation of Substances 2.0 (EUSES 2.0). Prepared for the European Chemicals Bureau by the National Institute of Public Health and the Environment (RIVM), Bilthoven. The Netherlands (RIVM Report no. 601900005). Available via the European Chemicals Bureau; 2004 [http://ecb.jrc.it]. EC. SCHER (Scientific Committee on Health and Environmental Risks), SCENIHR (Scientific Committee on Emerging and Newly Identified Health Risks), SCCS (Scientific Committee on Consumer Safety). Addressing the New Challenges for Risk Assessment, March 2013, European Commission; 2013. Fantke P, Ronnie Juraske, Assumpció A, Friedrich R, Jolliet O. Dynamic multicrop model to characterize impacts of pesticides in food. Environ Sci Technol 2011;45:8842–9. Gouin T, Thomas O, Cousins I, Barber J, Mackay D, Jones KC. Air-surface exchange of polybrominated diphenyl ethers and polychlorinated biphenyls. Environ Sci Technol 2002;36:1426–34. Gu Z, Feng J, Han W, Li L, Wu M, Fu J, et al. Diurnal variations of polycyclic aromatic hydrocarbons associated with PM2.5 in Shanghai, China. J Environ Sci 2010;22:389–96. Hornbuckle KC, Eisenreich SJ. Dynamics of gaseous semivolatile organic compounds in a terrestrial ecosystem—effects of diurnal and seasonal climate variations. Atmos Environ 1996;30:3935–45. Jurik TW. Temporal and spatial patterns of specific leaf weight in successional northern hardwood tree species. Am J Bot 1986;73:1083–92. Kishida M, Nisjikawa A, Fujimori K, Shibutani Y. Gas-particle concentrations of atmospheric polycyclic aromatic hydrocarbons at an urban and a residential site in Osaka, Japan: effect of formation of atmospherically stable layer on their temporal change. J Hazard Mater 2011;192:1340–9. Komprda J, Kubosova K, Dvorska A, Scheringer M, Klanova J, Holoubek I. Application of an unsteady state environmental distribution model to a decadal time series of PAH concentrations in Central Europe. J Environ Monitor 2009;11:269–76. Larcher W. Physiological plant ecology. 4th ed. Berlin: Springer; 2003. Larssen S, Sluyter R, Helmis C. Criteria for EUROAIRNET. The EEA Air Quality Monitoring and Information Network—Technical Report No. 12; 1999.

337

Lee RGM, Hung H, Mackay D, Jones KC. Measurement and modelling of the diurnal cycling of atmospheric PCBs and PAHs. Environ Sci Technol 1998;32:2172–9. Luo T, Li M, Luo J. Seasonal variations in leaf δ13C and nitrogen associated with foliage turnover and carbon gain for a wet subalpine fir forest in the Gongga Mountains, eastern Tibetan Plateau. Ecol Res 2011;26:253–63. McCrady JK. Vapor-phase 2,3,7,8-TCDD sorption to plant foliage—a species comparison. Chemosphere 1994;28:207–16. McLachlan MS, Horstmann M. Forest as filters of airborne organic pollutants: a model. Environ Sci Technol 1998;32:413–20. Morselli M, Ghirardello D, Semplice M, Di Guardo A. Modelling short-term varaibility of semivolatile organic chemicals in air at a local scale: an integrated modelling approach. Environ Pollut 2011;159:1406–12. Morville S, Delhomme O, Millet M. Seasonal and diurnal variations of atmospheric PAH concentrations between rural, suburban and urban areas. Atmos Pollut Res 2011;2: 363–73. Nasahara KN, Muraoka H, Nagai S, Mikami H. Vertical integration of leaf area index in a Japanese deciduous broad-leaved forest. Agr Forest Meteorol 2008;148:1136–46. Nizzetto L, Pastore C, Liu X, Camporini P, Stroppiana D, Herbert B, et al. Accumulation parameters and seasonal trend for PCBs in temperate and boreal forest plant species. Environ Sci Technol 2008;42:5911–6. Nizzetto L, Stroppiana D, Brivio PA, Boschetti M, Di Guardo A. Tracing the fate of PCBs in forest ecosystems. J Environ Monitor 2007;9:542–9. Nizzetto L, Jones KC, Gramatica O, Papa E, Cerabolini B, Di Guardo A. Accumulation of persistent organic pollutants in canopies of different forest type: role of species composition and altitudinal-temperature gradient. Environ Sci Technol 2006;40:6580–6. Paterson S, Mackay D, Gladman A. A fugacity model of chemical uptake by plants from soil and air. Chemosphere 1991;23:539–65. Poorter H, Evans JR. Photosynthetic nitrogen-use efficiency of species that differ inherently in specific leaf area. Oecologia 1998;116:26–37. Priemer DA, Diamond ML. Application of the Multimedia Urban Model to compare the fate of SOCs in an urban and forested watershed. Environ Sci Technol 2002;36: 1004–13. Rein A, Legind CN, Trapp S. New concepts for dynamic plant uptake models. SAR QSAR Environ Res 2011;22:191–215. REPA. Rapporto sulla qualità dell' aria di Como e provincia; 2007. REPA. Regional Environmental Protection Agency official web site; 2014 [http://ita. arpalombardia.it/ITA/qaria/doc_RichiestaDati.asp. Last access on April, 2014]. Saporta G. Probabilités, analyse des données et statistique. Paris: Éditions Technip; 1990. Simioni G, Gignoux J, Le Roux X, Appé R, Benest D. Spatial and temporal variations in leaf area index, specific leaf area and leaf nitrogen of two co-occurring savanna tree species. Tree Physiol 2004;24:205–16. Steyaert NLL, Hauck M, Van Hulle SWH, Hendriks AJ. Modelling bioaccumulation of semivolatile organic compounds (SOCs) from air in plants based on allometric principles. Chemosphere 2009;77:727–32. Terzaghi E, Morselli M, Semplice M, Vianelli A., Cerabolini BEL, Di Guardo A. Modelling the temporal uptake of PAHs in plants using a ecological realistic scenario. In preparation. Terzaghi E, Wild E, Zacchello G, Cerabolini BEL, Jones KC, Di Guardo A. Forest filter effect: role of leaves in capturing/releasing air particulate matter and its associated PAHs. Atmos Environ 2013;74:378–84. Trapp S, Mc Farlane C, Matthies M. Model for uptake of xenobiotics to plant: validation with bromacil experiments. Environ Toxicol Chem 1994;13:413–22. Undemann E, Czub G, McLachlan MS. Addressing temporal variability when modeling bioaccumulation in plants. Environ Sci Technol 2009;43:3751–6. Vile D, Garnier É, Shipley B, Laurent G, Navas M-L, Roumet C, et al. Specific leaf area and dry matter content estimate thickness in laminar leaves. Ann Botany 2005:1129–36. Wania F, McLachlan M. Estimating the influence of forest on the overall fate of SOCs using a multimedia fate model. Environ Sci Technol 2001;35:582–90. Wegmann F, Scheringer M, Muller M, Hungerbuhler K. Influence of vegetation on the environmental partitioning of DDT in two global multimedia models. Environ Sci Technol 2004;38:1505–12. Wesolowski T, Rowinski P. Timing of bud burst and tree-leaf development in a multispecies temperate forest. Forest Ecol Manag 2006;237:387–93. Wilcke W. Global patterns of polycyclic aromatic hydrocarbons (PAHs) in soil. Geoderma 2007;141:157–66.